1<html> 2<head> 3<title>pcre2posix specification</title> 4</head> 5<body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB"> 6<h1>pcre2posix man page</h1> 7<p> 8Return to the <a href="index.html">PCRE2 index page</a>. 9</p> 10<p> 11This page is part of the PCRE2 HTML documentation. It was generated 12automatically from the original man page. If there is any nonsense in it, 13please consult the man page, in case the conversion went wrong. 14<br> 15<ul> 16<li><a name="TOC1" href="#SEC1">SYNOPSIS</a> 17<li><a name="TOC2" href="#SEC2">DESCRIPTION</a> 18<li><a name="TOC3" href="#SEC3">COMPILING A PATTERN</a> 19<li><a name="TOC4" href="#SEC4">MATCHING NEWLINE CHARACTERS</a> 20<li><a name="TOC5" href="#SEC5">MATCHING A PATTERN</a> 21<li><a name="TOC6" href="#SEC6">ERROR MESSAGES</a> 22<li><a name="TOC7" href="#SEC7">MEMORY USAGE</a> 23<li><a name="TOC8" href="#SEC8">AUTHOR</a> 24<li><a name="TOC9" href="#SEC9">REVISION</a> 25</ul> 26<br><a name="SEC1" href="#TOC1">SYNOPSIS</a><br> 27<P> 28<b>#include <pcre2posix.h></b> 29</P> 30<P> 31<b>int regcomp(regex_t *<i>preg</i>, const char *<i>pattern</i>,</b> 32<b> int <i>cflags</i>);</b> 33<br> 34<br> 35<b>int regexec(const regex_t *<i>preg</i>, const char *<i>string</i>,</b> 36<b> size_t <i>nmatch</i>, regmatch_t <i>pmatch</i>[], int <i>eflags</i>);</b> 37<br> 38<br> 39<b>size_t regerror(int <i>errcode</i>, const regex_t *<i>preg</i>,</b> 40<b> char *<i>errbuf</i>, size_t <i>errbuf_size</i>);</b> 41<br> 42<br> 43<b>void regfree(regex_t *<i>preg</i>);</b> 44</P> 45<br><a name="SEC2" href="#TOC1">DESCRIPTION</a><br> 46<P> 47This set of functions provides a POSIX-style API for the PCRE2 regular 48expression 8-bit library. See the 49<a href="pcre2api.html"><b>pcre2api</b></a> 50documentation for a description of PCRE2's native API, which contains much 51additional functionality. There are no POSIX-style wrappers for PCRE2's 16-bit 52and 32-bit libraries. 53</P> 54<P> 55The functions described here are just wrapper functions that ultimately call 56the PCRE2 native API. Their prototypes are defined in the <b>pcre2posix.h</b> 57header file, and on Unix systems the library itself is called 58<b>libpcre2-posix.a</b>, so can be accessed by adding <b>-lpcre2-posix</b> to the 59command for linking an application that uses them. Because the POSIX functions 60call the native ones, it is also necessary to add <b>-lpcre2-8</b>. 61</P> 62<P> 63Those POSIX option bits that can reasonably be mapped to PCRE2 native options 64have been implemented. In addition, the option REG_EXTENDED is defined with the 65value zero. This has no effect, but since programs that are written to the 66POSIX interface often use it, this makes it easier to slot in PCRE2 as a 67replacement library. Other POSIX options are not even defined. 68</P> 69<P> 70There are also some options that are not defined by POSIX. These have been 71added at the request of users who want to make use of certain PCRE2-specific 72features via the POSIX calling interface or to add BSD or GNU functionality. 73</P> 74<P> 75When PCRE2 is called via these functions, it is only the API that is POSIX-like 76in style. The syntax and semantics of the regular expressions themselves are 77still those of Perl, subject to the setting of various PCRE2 options, as 78described below. "POSIX-like in style" means that the API approximates to the 79POSIX definition; it is not fully POSIX-compatible, and in multi-unit encoding 80domains it is probably even less compatible. 81</P> 82<P> 83The header for these functions is supplied as <b>pcre2posix.h</b> to avoid any 84potential clash with other POSIX libraries. It can, of course, be renamed or 85aliased as <b>regex.h</b>, which is the "correct" name. It provides two 86structure types, <i>regex_t</i> for compiled internal forms, and 87<i>regmatch_t</i> for returning captured substrings. It also defines some 88constants whose names start with "REG_"; these are used for setting options and 89identifying error codes. 90</P> 91<br><a name="SEC3" href="#TOC1">COMPILING A PATTERN</a><br> 92<P> 93The function <b>regcomp()</b> is called to compile a pattern into an 94internal form. By default, the pattern is a C string terminated by a binary 95zero (but see REG_PEND below). The <i>preg</i> argument is a pointer to a 96<b>regex_t</b> structure that is used as a base for storing information about 97the compiled regular expression. (It is also used for input when REG_PEND is 98set.) 99</P> 100<P> 101The argument <i>cflags</i> is either zero, or contains one or more of the bits 102defined by the following macros: 103<pre> 104 REG_DOTALL 105</pre> 106The PCRE2_DOTALL option is set when the regular expression is passed for 107compilation to the native function. Note that REG_DOTALL is not part of the 108POSIX standard. 109<pre> 110 REG_ICASE 111</pre> 112The PCRE2_CASELESS option is set when the regular expression is passed for 113compilation to the native function. 114<pre> 115 REG_NEWLINE 116</pre> 117The PCRE2_MULTILINE option is set when the regular expression is passed for 118compilation to the native function. Note that this does <i>not</i> mimic the 119defined POSIX behaviour for REG_NEWLINE (see the following section). 120<pre> 121 REG_NOSPEC 122</pre> 123The PCRE2_LITERAL option is set when the regular expression is passed for 124compilation to the native function. This disables all meta characters in the 125pattern, causing it to be treated as a literal string. The only other options 126that are allowed with REG_NOSPEC are REG_ICASE, REG_NOSUB, REG_PEND, and 127REG_UTF. Note that REG_NOSPEC is not part of the POSIX standard. 128<pre> 129 REG_NOSUB 130</pre> 131When a pattern that is compiled with this flag is passed to <b>regexec()</b> for 132matching, the <i>nmatch</i> and <i>pmatch</i> arguments are ignored, and no 133captured strings are returned. Versions of the PCRE library prior to 10.22 used 134to set the PCRE2_NO_AUTO_CAPTURE compile option, but this no longer happens 135because it disables the use of backreferences. 136<pre> 137 REG_PEND 138</pre> 139If this option is set, the <b>reg_endp</b> field in the <i>preg</i> structure 140(which has the type const char *) must be set to point to the character beyond 141the end of the pattern before calling <b>regcomp()</b>. The pattern itself may 142now contain binary zeros, which are treated as data characters. Without 143REG_PEND, a binary zero terminates the pattern and the <b>re_endp</b> field is 144ignored. This is a GNU extension to the POSIX standard and should be used with 145caution in software intended to be portable to other systems. 146<pre> 147 REG_UCP 148</pre> 149The PCRE2_UCP option is set when the regular expression is passed for 150compilation to the native function. This causes PCRE2 to use Unicode properties 151when matchine \d, \w, etc., instead of just recognizing ASCII values. Note 152that REG_UCP is not part of the POSIX standard. 153<pre> 154 REG_UNGREEDY 155</pre> 156The PCRE2_UNGREEDY option is set when the regular expression is passed for 157compilation to the native function. Note that REG_UNGREEDY is not part of the 158POSIX standard. 159<pre> 160 REG_UTF 161</pre> 162The PCRE2_UTF option is set when the regular expression is passed for 163compilation to the native function. This causes the pattern itself and all data 164strings used for matching it to be treated as UTF-8 strings. Note that REG_UTF 165is not part of the POSIX standard. 166</P> 167<P> 168In the absence of these flags, no options are passed to the native function. 169This means the the regex is compiled with PCRE2 default semantics. In 170particular, the way it handles newline characters in the subject string is the 171Perl way, not the POSIX way. Note that setting PCRE2_MULTILINE has only 172<i>some</i> of the effects specified for REG_NEWLINE. It does not affect the way 173newlines are matched by the dot metacharacter (they are not) or by a negative 174class such as [^a] (they are). 175</P> 176<P> 177The yield of <b>regcomp()</b> is zero on success, and non-zero otherwise. The 178<i>preg</i> structure is filled in on success, and one other member of the 179structure (as well as <i>re_endp</i>) is public: <i>re_nsub</i> contains the 180number of capturing subpatterns in the regular expression. Various error codes 181are defined in the header file. 182</P> 183<P> 184NOTE: If the yield of <b>regcomp()</b> is non-zero, you must not attempt to 185use the contents of the <i>preg</i> structure. If, for example, you pass it to 186<b>regexec()</b>, the result is undefined and your program is likely to crash. 187</P> 188<br><a name="SEC4" href="#TOC1">MATCHING NEWLINE CHARACTERS</a><br> 189<P> 190This area is not simple, because POSIX and Perl take different views of things. 191It is not possible to get PCRE2 to obey POSIX semantics, but then PCRE2 was 192never intended to be a POSIX engine. The following table lists the different 193possibilities for matching newline characters in Perl and PCRE2: 194<pre> 195 Default Change with 196 197 . matches newline no PCRE2_DOTALL 198 newline matches [^a] yes not changeable 199 $ matches \n at end yes PCRE2_DOLLAR_ENDONLY 200 $ matches \n in middle no PCRE2_MULTILINE 201 ^ matches \n in middle no PCRE2_MULTILINE 202</pre> 203This is the equivalent table for a POSIX-compatible pattern matcher: 204<pre> 205 Default Change with 206 207 . matches newline yes REG_NEWLINE 208 newline matches [^a] yes REG_NEWLINE 209 $ matches \n at end no REG_NEWLINE 210 $ matches \n in middle no REG_NEWLINE 211 ^ matches \n in middle no REG_NEWLINE 212</pre> 213This behaviour is not what happens when PCRE2 is called via its POSIX 214API. By default, PCRE2's behaviour is the same as Perl's, except that there is 215no equivalent for PCRE2_DOLLAR_ENDONLY in Perl. In both PCRE2 and Perl, there 216is no way to stop newline from matching [^a]. 217</P> 218<P> 219Default POSIX newline handling can be obtained by setting PCRE2_DOTALL and 220PCRE2_DOLLAR_ENDONLY when calling <b>pcre2_compile()</b> directly, but there is 221no way to make PCRE2 behave exactly as for the REG_NEWLINE action. When using 222the POSIX API, passing REG_NEWLINE to PCRE2's <b>regcomp()</b> function 223causes PCRE2_MULTILINE to be passed to <b>pcre2_compile()</b>, and REG_DOTALL 224passes PCRE2_DOTALL. There is no way to pass PCRE2_DOLLAR_ENDONLY. 225</P> 226<br><a name="SEC5" href="#TOC1">MATCHING A PATTERN</a><br> 227<P> 228The function <b>regexec()</b> is called to match a compiled pattern <i>preg</i> 229against a given <i>string</i>, which is by default terminated by a zero byte 230(but see REG_STARTEND below), subject to the options in <i>eflags</i>. These can 231be: 232<pre> 233 REG_NOTBOL 234</pre> 235The PCRE2_NOTBOL option is set when calling the underlying PCRE2 matching 236function. 237<pre> 238 REG_NOTEMPTY 239</pre> 240The PCRE2_NOTEMPTY option is set when calling the underlying PCRE2 matching 241function. Note that REG_NOTEMPTY is not part of the POSIX standard. However, 242setting this option can give more POSIX-like behaviour in some situations. 243<pre> 244 REG_NOTEOL 245</pre> 246The PCRE2_NOTEOL option is set when calling the underlying PCRE2 matching 247function. 248<pre> 249 REG_STARTEND 250</pre> 251When this option is set, the subject string starts at <i>string</i> + 252<i>pmatch[0].rm_so</i> and ends at <i>string</i> + <i>pmatch[0].rm_eo</i>, which 253should point to the first character beyond the string. There may be binary 254zeros within the subject string, and indeed, using REG_STARTEND is the only 255way to pass a subject string that contains a binary zero. 256</P> 257<P> 258Whatever the value of <i>pmatch[0].rm_so</i>, the offsets of the matched string 259and any captured substrings are still given relative to the start of 260<i>string</i> itself. (Before PCRE2 release 10.30 these were given relative to 261<i>string</i> + <i>pmatch[0].rm_so</i>, but this differs from other 262implementations.) 263</P> 264<P> 265This is a BSD extension, compatible with but not specified by IEEE Standard 2661003.2 (POSIX.2), and should be used with caution in software intended to be 267portable to other systems. Note that a non-zero <i>rm_so</i> does not imply 268REG_NOTBOL; REG_STARTEND affects only the location and length of the string, 269not how it is matched. Setting REG_STARTEND and passing <i>pmatch</i> as NULL 270are mutually exclusive; the error REG_INVARG is returned. 271</P> 272<P> 273If the pattern was compiled with the REG_NOSUB flag, no data about any matched 274strings is returned. The <i>nmatch</i> and <i>pmatch</i> arguments of 275<b>regexec()</b> are ignored (except possibly as input for REG_STARTEND). 276</P> 277<P> 278The value of <i>nmatch</i> may be zero, and the value <i>pmatch</i> may be NULL 279(unless REG_STARTEND is set); in both these cases no data about any matched 280strings is returned. 281</P> 282<P> 283Otherwise, the portion of the string that was matched, and also any captured 284substrings, are returned via the <i>pmatch</i> argument, which points to an 285array of <i>nmatch</i> structures of type <i>regmatch_t</i>, containing the 286members <i>rm_so</i> and <i>rm_eo</i>. These contain the byte offset to the first 287character of each substring and the offset to the first character after the end 288of each substring, respectively. The 0th element of the vector relates to the 289entire portion of <i>string</i> that was matched; subsequent elements relate to 290the capturing subpatterns of the regular expression. Unused entries in the 291array have both structure members set to -1. 292</P> 293<P> 294A successful match yields a zero return; various error codes are defined in the 295header file, of which REG_NOMATCH is the "expected" failure code. 296</P> 297<br><a name="SEC6" href="#TOC1">ERROR MESSAGES</a><br> 298<P> 299The <b>regerror()</b> function maps a non-zero errorcode from either 300<b>regcomp()</b> or <b>regexec()</b> to a printable message. If <i>preg</i> is not 301NULL, the error should have arisen from the use of that structure. A message 302terminated by a binary zero is placed in <i>errbuf</i>. If the buffer is too 303short, only the first <i>errbuf_size</i> - 1 characters of the error message are 304used. The yield of the function is the size of buffer needed to hold the whole 305message, including the terminating zero. This value is greater than 306<i>errbuf_size</i> if the message was truncated. 307</P> 308<br><a name="SEC7" href="#TOC1">MEMORY USAGE</a><br> 309<P> 310Compiling a regular expression causes memory to be allocated and associated 311with the <i>preg</i> structure. The function <b>regfree()</b> frees all such 312memory, after which <i>preg</i> may no longer be used as a compiled expression. 313</P> 314<br><a name="SEC8" href="#TOC1">AUTHOR</a><br> 315<P> 316Philip Hazel 317<br> 318University Computing Service 319<br> 320Cambridge, England. 321<br> 322</P> 323<br><a name="SEC9" href="#TOC1">REVISION</a><br> 324<P> 325Last updated: 15 June 2017 326<br> 327Copyright © 1997-2017 University of Cambridge. 328<br> 329<p> 330Return to the <a href="index.html">PCRE2 index page</a>. 331</p> 332